Table 1.

Summary of features extracted from data series generated by each sensor (x) per each window of size N. The frequency-domain features were calculated using the Fourier Transform coefficients ($F_i$).

Domain	No.	Name	Equation	Description	References
Time	1	Mean	$\mu \left(x\right)=\frac{1}{N}{\sum }_{i=1}^N{x}_i$	Average of the data	[15,19,20,25,26,27,28,29]
	2	Range	$r\left(x\right)=max\left(x\right)-min\left(x\right)$	Difference between the greatest and the smallest values in the data	[29]
	3	Standard Deviation	$\sigma \left(x\right)=\sqrt{\frac{1}{N}{\sum }_{i=1}^N{(x_i-\mu )}^2}$	Measure of dispersion in the data	[15,20,25,27,29]
	4	Skewness	$s\left(x\right)=\frac{1}{N{\sigma }^3}{\sum }_{i=1}^N{(x_i-\mu )}^3$	Measure of asymmetry of a distribution around its mean	[25,26,28]
	5	Kurtosis	$k\left(x\right)=\frac{1}{N{\sigma }^4}{\sum }_{i=1}^N{(x_i-\mu )}^4$	Measure of how different a distribution’s tails are from the tails of a normal distribution	[25,26,28]
Frequency	6	DFR	$DFR\left(x\right)=\frac{max\left(\{F_1,...,F_n\}\right)}{{\displaystyle \sum _{i=1}^{N/2}}{F}_i}$	Dominant Frequency Ratio, ratio of highest magnitude FFT coefficient to the sum of magnitudes of all FFT coefficients	[29]
	7	Entropy	$entropy\left(x\right)=-{\displaystyle \sum _{i=1}^{N/2}}{F}_i{log}_2{F}_i$	Information entropy of the normalised values of FFT coefficient magnitude	[15,29]
	8	Energy	$energy\left(x\right)={\displaystyle \sum _{i=1}^{N/2}}{F}_i^2$	Sum of the squared discrete FFT component magnitudes	[19,25,28]